INDEX

Explanations

Code and metadata

np_max-act · gemini-2.0-flash

New Auto-Interp

Configuration

mwhanna/qwen3-4b-transcoders/layer_23.safetensors

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

163,840

Data Type

float32

Hook Name

blocks.23.mlp.hook_in

Architecture

transcoder

Context Size

8,192

Dataset

monology/pile-uncopyrighted

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 affairs

-0.27

 trade

-0.26

.exports

-0.25

åĲģ

-0.25

 Fragment

-0.24

.exclude

-0.24

çĶŁæ´»

-0.24

trade

-0.23

ivers

-0.23

å¹²

-0.23

POSITIVE LOGITS

prep

0.27

diÄŁi

0.26

"+"

0.25

ouri

0.25

ç»¿

0.25

Ä±s

0.24

ntl

0.24

icial

0.24

 rootView

0.24

altura

0.24

Activations Density 0.001%

Code and metadata

No Comments

No Known Activations